Query expansion based on Conceptual Word Cluster Space Graph

For the information retrieval problem, the query expansion methods have been studied for a long time. However, the performance still can´t meet the requirements of the users. In this paper, we present a directed graph model, named Conceptual Word Cluster Space Graph (CWCSG), which is constructed to express the semantic similarity among concepts. There are three steps to construct the CWCSG. First, the conceptual words are transformed to different terms. Then, the terms are clustered to different Conceptual Word Clusters (CWCs). After the extraction of CWCs, we calculate the co-occurrence relationship between CWCs to construct CWCSG. Based on CWCSG, user´s query is extended to meet user´s search requirement in a more accurate way. The experiments show that our method can provide a better performance comparing to classic synonymous dictionary of WordNet.